Purity Shmurity

It is often said or implied that a language is in danger if it is borrowing too many words from others, becoming too influenced by others etc. Witness Russian, Arab and French delusions regarding Anglicisms and English code-mixing. Or Nazi anxieties that the "mongrelization" that befell Yiddish and English might one day afflict German and debase its Germanic purity.

The thing is though, that examination of linguistic history will demonstrate that the exact opposite is the case. The more open a language and its speakers are to outside influence, to borrowing, to loanwords and code-mixing, the more viable it is.

Thanks to 19th century Gallo-Russian and Germano-Russian bilingualism in the upper classes, it is impossible to write an extended passage in standard literary Russian today without French influence being evident in every other sentence, and German influence in every paragraph. Russian did not borrow an extreme amount of lexical items (as English did) so much as duplicate French morphological and phrasal patterns, and some aspects of German sentence structure. Morphological and phrasal calques are legion, and in fact probably make up around half of the contents of your average Russian dictionary. There were intentionally gallicizing texts that were written in the 19th century, sometimes in order to satirize hyper-gallicism, and at other times in order to signal that e.g. a fictional character should be understood as speaking in French even if the words are in Russian. Today many such texts do not read as foreignized at all. Some, like Tatyana's Letter to Onegin in Pushkin's verse novel, just read as beautiful Russian.

Of course, Russian would have continued to exist just fine without its various (overlapping) episodes of Church Slavonic, Gallic and Germanic influence, as well as the oft-denied period of extensive Polish influence in the late middle ages, and the oft-decried period of English influence today. But it is certainly the richer for all of them.

In other cases, though, being open to foreign influence is a matter of language life and language death. Particularly in situations of prolonged asymmetrical bilingualism. (i.e. when it becomes a "Minority Language" like Modern Occitan, Modern Catalan or Modern Welsh.)

Minority bilingual groups that regard foreign influence of the majority language as something to be avoided at all costs, as a sort of ethnic sin for which penance is required, tend in fact to be the ones whose languages die most quickly. If you can't say it "correctly" enough, or if there's no way to say it "correctly" without socially unacceptable calques and loanwords, then you have all the more reason to just switch completely to the majority language. Lexical purism is not as much of an issue, particularly in an age when digital devices can allow easy lookup. On the other hand, if you're puristic about *syntactic* influence from the majority language, then you're either being unrealistic or you're dooming your own language out of reverence for it.

This problem of purism is one that has exacerbated the situation and lowered the prospects for many endangered Native American languages, even in the face of wide-spread interest in their survival.
I might wager that this has something to do with why Aramaic-speaking Jewish communities in early medieval Europe so quickly shifted to (lexically Judaized) European languages. Not "despite the sanctity of Aramaic" but because of it. As opposed to the easy survival of other Jewish languages....well, before the Holocaust anyway (and before the re-vernacularization of Hebrew.)
If you're okay with borrowings in syntax, then stable bilingualism is far easier to maintain.

Romani has developed (almost) unique devices for the incorporation of loanwords, maintaining a boundary between Them and Us by using different morphology (and in many dialects a somewhat different stress pattern) for European loans. Time and again linguists have remarked on the unusually profound and widespread degree of borrowing in Romani dialects at all levels. If borrowing were the way toward language death, then Romani would have died out hundreds of years ago, as indeed did most of the other once-numerous* diasporic Indic languages throughout Eurasia, of which most have left only traces by which to infer their past existence, and of which only a handful of others have survived long enough to be documented.  Romani, though, thrived. Contrary to what one might have predicted. Some dialects of it died, to be sure, and many others are endangered. But much of this is quite recent. Many dialects of it are doing quite well. (It has vastly more speakers than many territorial European languages. More than Maltese, Icelandic and Slovenian put together.) It remained vigorous for centuries in contexts where any sociolinguist might have reasonably guessed that it wouldn't be long for this world. In part because it borrowed the living shit out of co-territorial languages. It made linguistic boundaries more permeable, even as it made them all the more manifest. I wouldn't go so far as to say that the tenacity of Romani in surviving for so long or its level of syntactic, morphological and lexical borrowing are unique. They aren't. But they're certainly remarkable. Romani, no less than the Roma, survived by sheer adaptability.

Promiscuity is often the way of survival. Life-long virgins are not known for their many grandchildren.

If you're wondering why they were so common — Northward migration, usually temporary but obviously in a few cases permanent, of all sorts of trade-based communities and kin-groups appears to have been commonplace in ancient and medieval India. Usually for economic reasons. Documentation of trade-based Indian communities outside of India goes back to Roman times. The origin of the Roma is only a mystery in popular imagination and in the minds of those who want a sexier answer.


  1. Between the theoretical principle of one-phoneme-per-letter and the “field” practice of two-phonemes-per-letter lies the little gap that separates a pure phoneme from a hybrid phoneme. A hybrid phoneme is (at least) two phonemes in one: some of its allophones belong to one language, some to another. A hybrid phoneme is born when two sounds that are “phonetically similar” mate and fuse into one sound. This hybrid sound can be conceptualized in theory as either the union or the intersection of the two sounds: as either containing both its “parent” phonemes, and therefore belonging naturally to both languages, or else as an illegitimate half-breed bastard, born in “the field,” wandering in the structural no-man’s-land between the two (legitimate) parent languages, but belonging to neither. A pidgin, creole, mestizo phoneme: a source of pollution, noise, interference. An undocumented alien sound illegally crossing the linguistic border to “pass” for a legitimate sound of a language to which it doesn’t “really” belong.
    For how we imagine hybrid phonemes may depend, in the final analysis, on our attitudes to purity and pollution. These attitudes shape the way the topic of hybrid phonemes is, implicitly or explicitly, conceptualized and talked about —or passed over in silence. If we think that pure is better than hybrid, then we will tend to think of pure phonemes as “occurring in nature,” gathered in the field by scientists, and we will tend to think of hybrid phonemes as unnatural noise, pollution, a symptom of “outside influences,” of foreign powers acting at a distance. Normally, people don’t stretch a phoneme into a second language, because “normal” people are monoglot natives.
    If, on the other hand, we think that normal people speak more than one language, we will tend to think that hybrid phonemes thrive in their natural habitat, in the speech of so-called “bilinguals.” — Tomás Kalmar _Illegal Alphabets_ Epilogue.